Association of the Indoor Environment With Dry Eye Metrics

This cross-sectional study assesses the association between the indoor environment and the symptoms and signs of dry eye in a sample of US veterans.

Key Points

Question

Are there associations between the indoor environment, specifically temperature, humidity, and air pollutants, and the symptoms and signs of dry eye?

Findings

In this cross-sectional study of 97 US veterans, humidity was positively associated with dry eye metrics. In multivariate models, concentration of particulate air pollutants was associated with dry eye symptoms and signs.

Meaning

These findings suggest that indoor environmental manipulations, such as regulating humidity and reducing airborne particulate matter, may be a therapeutic target in some individuals with dry eye.

Abstract

Importance

The ocular surface is continuously exposed to the environment. Although studies have focused on associations between outdoor environmental conditions and dry eye, information on associations between the indoor environment and dry eye is lacking.

Objective

To determine associations between the indoor environment and dry eye.

Design, Setting, and Participants

This prospective cross-sectional study sample of 97 veterans with a wide range of dry eye metrics was recruited from the Miami Veterans Affairs Healthcare eye clinic from October 19, 2017, to August 30, 2018. Dry eye metrics were first evaluated in the clinic, followed by indoor home environmental metrics within 1 week using a handheld particle counter. Data were analyzed from October 19, 2017, to August 30, 2018.

Main Outcomes and Measures

Symptoms of dry eye were assessed with standardized questionnaires. Dry eye signs were assessed via standard examination. Indoor environmental metrics included temperature, humidity, and particulate matter mass and count.

Results

Of the 97 participants included in the analysis, 81 (84%) were men, with a mean (SD) age of 58.2 (11.9) years. Dry eye symptoms were in the moderate range with a mean (SD) Ocular Surface Disease Index (OSDI) score of 31.2 (23.6). Humidity was associated with worse symptoms and signs, including OSDI score (r = 0.30 [95% CI, 0.07-0.49]; P = .01), inflammation (r = 0.32 [95% CI, 0.10-0.51]; P = .01), Schirmer score (r = −0.25 [95% CI, −0.45 to 0.02]; P = .03), eyelid vascularity (r = 0.27 [95% CI, 0.05-0.47]; P = .02), and meibomian gland dropout (r = 0.27 [95% CI, 0.05-0.47]; P = .02). In multivariate analyses, particulate matter of 2.5 μm or less (PM_2.5) was associated with dry eye metrics when adjusted for demographic characteristics, comorbidities, medications, and interaction variables. For example, a 1-unit increase in instrumented PM_2.5 level was associated with a 1.59 increase in the OSDI score (95% CI, 0.58-2.59; P = .002), a 0.39 reduction in Schirmer score (95% CI, −0.75 to −0.03; P = .04), a 0.07 increase in meibomian gland dropout (95% CI, 0.01-0.13; P = .02), and a 0.06 increase in inflammation (95% CI, 0.02-0.11; P = .009).

Conclusions and Relevance

When adjusting for humidity, this study found that increased particulate matter exposure was associated with worse dry eye metrics. Humidity was positively associated with dry eye metrics, potentially because higher humidity increases microbial growth and particulate matter size and mass.

Introduction

Dry eye is a multifactorial chronic disease of the tears and ocular surface.¹ Worldwide, the prevalence of dry eye ranges from 5% to 50% and varies owing to population characteristics, disease definition, and other risk factors.² In the United States, approximately 16.4 million adults (6.8%) have been diagnosed with dry eye.³ Symptoms of dry eye are heterogeneous and include painful symptoms (ie, dryness, burning, irritation) and vision-related symptoms (ie, poor or fluctuating vision), which can negatively affect physical health, mental health, and quality of life.² Signs of dry eye are likewise heterogeneous and include decreased tear production, increased tear evaporation, inflammation, and high tear osmolarity.⁴ Numerous risk factors for dry eye have been identified, including age, sex, comorbidities such as depression and arthritis, and medications such as antihistamines.²

Dry eye is also influenced by environmental factors, including wind,⁵ high temperature,⁶ low humidity,⁵ high altitude,⁷ and air pollution.⁵ In a study of 500 hospital-based individuals in India,⁶ wind and high temperature exposure, as assessed by self-report, correlated with dry eye prevalence (odds ratios, 2.15 and 1.91, respectively). In a study of 3.41 million US veterans, the risk of a diagnosis of dry eye was 6% and 7% higher in zip codes where wind speed and humidity were 1 SD less than the mean, respectively. The risk of a dry eye diagnosis was also 13% higher in zip codes where aerosol optical depth, a measure of atmospheric aerosols that include airborne particulate matter as well as natural aerosols (such as water vapors), was 1 SD greater than the mean.⁵ Incidence of dry eye has also been reported to increase during spring, mirroring the time of highest pollen levels.⁸ Adding support to the connection between particulate matter and dry eye, particulate matter exposure has been connected to inflammatory injury and oxidative damage,^9,10,11 2 biological mechanisms relevant in dry eye.¹²

Most environmental studies, including those described above, examined associations between the ambient (or outdoor) environment and dry eye. Fewer studies have focused on the effect of indoor environmental conditions (including indoor air pollutants such as particulate matter) on dry eye, although humans spend 90% of their time indoors.¹³ In a study of 120 women in India,¹⁴ the number of indoor air pollution sources (including kerosene fuel, smoke, and vehicular exhaust) was correlated with self-reported eye irritation (ρ = 0.56; P < .05). In another cross-sectional study of 3335 employees in Japan,¹⁵ low indoor humidity and cold temperature as measured by self-administered questionnaires were correlated with the presence of eye irritation (odds ratios, 1.28 and 1.61, respectively). However, these studies assessed indoor conditions by self-report and not by direct measurements. To address this gap, this research examined associations between objectively measured indoor environmental metrics, namely temperature, humidity, and particulate matter mass concentration and count, and dry eye symptoms and signs (referred to as dry eye metrics hereinafter).

Methods

Study Population

South Florida residents with healthy eyelid and corneal anatomy were recruited from the Miami Veterans Affairs Healthcare eye clinic from October 19, 2017, to August 30, 2018. Inclusion was limited to patients who did not spend significant time away from home before or during the data collection period. Patients were excluded from participation if they had ocular or systemic conditions that could confound dry eye, including contact lens use, a history of refractive surgery, use of ocular medications with the exception of artificial tears, an active external ocular process, cataract surgery in the last 6 months, or a history of glaucoma or retinal surgery.¹⁶ Systemically, individuals with a diagnosis of HIV, sarcoidosis, Sjögren syndrome, graft-vs-host disease, or a collagen vascular disease were excluded. Patients with seasonal allergies were not excluded. All patients who visited the eye clinic and met inclusion and exclusion criteria were offered participation in the study regardless of whether they had dry eye symptoms or diagnosis. Interested patients were scheduled for a study visit in which written informed consent was obtained and dry eye metrics were measured. Patients received $100 as compensation for their participation in the study. The study was approved by the Miami Veterans Affairs Healthcare and University of Miami institutional review boards. This study was conducted in accordance with the principles of the Declaration of Helsinki¹⁷ and complied with the requirements of the US Health Insurance Portability and Accountability Act.

Questionnaires

For each participant, demographic data (age, sex, race, and ethnicity), ocular and medical history, and medication data were collected. Patients answered standardized questionnaires regarding symptoms of dry eye, including the Dry Eye Questionnaire 5 (scores range from 0 to 22, with higher scores indicating greater severity)¹⁸ and the Ocular Surface Disease Index (OSDI; scores range from 0 to 100, with higher scores indicating greater presence of disease).¹⁹

Ocular Surface Evaluation

All patients underwent a standardized ocular surface examination by an investigator masked to the questionnaire data, which included measurement of (1) tear osmolarity (TearLAB Osmolarity System; TearLAB) (once in each eye); (2) inflammation (InflammaDry; Quidel Corporation); (3) tear evaporation measured via tear breakup time (TBUT) (5 μL of fluorescein instilled in the superior conjunctivae, seconds measured until the first black spot appeared in the tear film, 3 measurements taken with 5-second blink interval between measurements, and the mean calculated); (4) corneal epithelial cell disruption measured with the National Eye Institute scale (5 areas of cornea assessed; each scored 0-3, for a possible total of 15, with higher scores indicating greater disruption); (5) Schirmer score with anesthesia measured as millimeters of wetting at 5 minutes; (6) eyelid parameters, including eyelid vascularity (0 indicates none; 1, mild; 2, moderate; and 3, severe); meibomian gland atrophy graded to the Meiboscale (by comparing with standardized pictures); and meibum quality (0 indicates clear liquid; 1, white liquid; 2, granular; 3, viscosity of toothpaste; and 4, no visible meibum extracted). For each individual, data from the more severely affected eye (lower value for TBUT and Schirmer score and higher value for staining and osmolarity) were used in the analyses.

Indoor Environmental Monitoring

Although the term indoor environment refers to various metrics, in the context of this study it refers to humidity, temperature, and airborne particulate matter (of different sizes) in participants’ homes. Particulate matter can be categorized in various ways, including by particle size, morphology, and chemical composition. In this study, we specifically focused on airborne particulate matter mass by 2 sizes, namely 2.5 μm or less (PM_2.5) and 10.0 μm or less (PM₁₀) in aerodynamic diameter, and counts of particles at least 0.5 μm and at least 5.0 μm in aerodynamic diameter.

Indoor environmental conditions were monitored in the home within 7 days of the clinic visit on average. A handheld particle counter (Aerocet 531; MetOne Instruments, Inc) was deployed for 90 minutes within 4 ft of the home’s air conditioning closet (or air handling unit) or within 8 ft of an air vent while the air conditioner was running.²⁰ For the first 45 minutes, the particle counter was set to mass mode, which estimated particulate matter mass (micrograms per cubic meter of air) by 4 sizes: 1.0, 2.5, 7.0, and 10.0 μm (PM₁, PM_2.5, PM₇, and PM₁₀, respectively). For the following 45 minutes, the particle counter was set to count mode, which counted the number of airborne particles measuring at least 0.5 μm and at least 5.0 μm/ft³ of air. Information on indoor temperature and humidity were collected by the particle counter simultaneously. Owing to technical difficulties, investigators were unable to record both count and mass data for some participants, resulting in number discrepancies.

Statistical Analysis

Data were analyzed from October 19, 2017, to August 30, 2018. Descriptive analysis was conducted to assess patients’ demographic characteristics, comorbidities, medication use, and dry eye and environmental measures. Correlation coefficients were calculated to evaluate associations between indoor environmental and dry eye metrics. Given that meteorological conditions can affect particles (eg, increased humidity can inflate their size owing to their hygroscopic nature, and elevated temperature can facilitate their dispersion), particulate matter cannot be considered an independent (or exogenous) variable. We thus used particulate matter as an endogenous variable and temperature and humidity as exogenous variables.²¹ Thus, particulate matter was instrumented on temperature and humidity separately in the multivariate analyses controlling for potential confounders, such as demographics and comorbidities. All analyses were conducted in SPSS, version 24.0 (IBM Corporation), and Stata, version 14.2 (StataCorp LLC). P values were 2-sided and not adjusted for multiple analyses performed; P < .05 indicated significance.

Results

Study Population and Indoor Environmental Metrics

The mean (SD) age of the 97 study participants included in the analysis was 58.2 (11.9) years. Eighty-one participants (84%) were men and 16 (16%) were women, 54 (56%) self-identified as black, and 65 (67%) self-identified as non-Hispanic (Table 1). Symptoms of dry eye were in the moderate range, with a mean (SD) Dry Eye Questionnaire 5 score of 10.5 (5.3) and a mean (SD) OSDI score of 31.2 (23.6) (Table 1). With regard to indoor environmental metrics, the mean (SD) temperature was 24.1 °C (2.1 °C), and the mean (SD) humidity was 52.4% (8.3%). Both the number of particles greater than 0.5 μm/ft³ of air and greater than 5 μm/ft³ of air varied between homes (114.4-5 300 000.0 μm/ft³ and 0-8843.3 μm/ft³, respectively), suggesting a skewed distribution. Mass concentrations of PM_2.5 and PM₁₀ (0-5.5 μg/m³ and 0-54.0 μg/m³, respectively) also varied between homes (Table 2). However, temperature, humidity, PM_2.5, and PM₁₀ did not vary by season.

Table 1. Clinical Characteristics of Study Population.

Variable	Patient dataa
Age, mean (SD), y	58.2 (11.9)
Male	81/97 (84)
White	43/97 (44)
Hispanic	32/97 (33)
Smoking
Past	48/97 (49)
Current	34/97 (35)
Dry eye symptoms	54/97 (56)
DES (ICD-10)	60/97 (62)
Comorbidities
Depression	62/96 (65)
Osteoarthritis	49/95 (52)
Hypercholesteremia	47/96 (49)
Hypertension	46/96 (48)
Diabetes	33/96 (34)
Sleep apnea	28/96 (29)
Posttraumatic stress disorder	21/96 (22)
Benign prostatic hyperplasia	18/96 (19)
Hepatitis C virus	15/96 (16)
Traumatic brain injury	4/96 (4)
Medications used
Analgesics	61/96 (64)
Antianxiety	44/96 (46)
Antidepressant	43/96 (45)
Cholesterol level–lowering agent	42/96 (44)
Nonsteroidal anti-inflammatory agent	26/96 (27)
Acetylsalicylic acid	26/96 (27)
β-Blocker	15/96 (16)
Dry eye symptoms, mean (SD)
DEQ-5 scoreb	10.5 (5.3)
OSDI scorec	31.2 (23.6)
Intensity of ocular pain for 1-wk recalld	2.7 (2.6)
Total NSPI-E scoree	17.6 (19.5)
Dry eye signs, mean (SD) [No. of patients]
Osmolarity, mOsm/Lf	319.3 (20.1) [96]
Inflammationg	1.2 (1.0) [94]
Tear breakup time, sh	5.9 (3.0) [94]
Corneal stainingi	1.7 (2.3) [95]
Schirmer score, mm of wetting at 5 minj	13.5 (7.8) [94]
Eyelid vascularityk	0.8 (1.1) [94]
Meibomian gland dropoutl	1.7 (1.2) [94]
Meibum qualitym	2.2 (1.3) [92]

Abbreviations: DEQ-5, Dry Eye Questionnaire 5; DES, dry eye syndrome; ICD-10, International Statistical Classification of Diseases and Related Health Problems, Tenth Revision; NSPI-E, Neuropathic Pain Symptom Inventory modified for the Eye; OSDI, Ocular Surface Disease Index.

^aUnless otherwise indicated, data are expressed as number/total number (percentage) of patients.

^bScores range from 0 to 22, with higher scores indicating greater severity (<6 indicates reference).¹⁸

^cScores range from 0 to 100, with higher scores indicating greater presence of disease (<12 indicates reference).¹⁹

^dScores range from 0 to 10, with higher scores indicating greater intensity of pain.

^eScores range from 0 to 100, with higher scores indicating more severe neuropathic ocular pain symptoms.

^fMeasured using the TearLAB Osmolarity System (TearLAB), with less than 316 mOsm/L as reference.

^gMeasured using InflammaDry (Quidel Corporation), with 0 indicating no pink stripe; 1, light pink stripe; 2, pink stripe; and 3, fuchsia stripe.

^hMeasured using 5 μL of fluorescein instilled in the superior conjunctivae as seconds measured until the first black spot appeared in the tear film; 3 measurements were taken with 5-second blink interval between measurements, and the mean was calculated. Greater than 10 seconds indicated reference.

ⁱScores range from 0 to 3, with higher scores indicating greater disruption.

^jLess than 10 mm indicated reference.

^kScores range from 0 to 3, with 0 indicating none; 1, mild; 2, moderate; and 3, severe.

^lScores range from 0 to 4, with higher scores indicating more severe gland dropout, based on standardized pictures provided by the Meiboscale.

^m Scores range from 0 to 4, with 0 indicating clear liquid; 1, white liquid; 2, granular; 3, viscosity of toothpaste; and 4, no visible meibum extracted.

Table 2. Indoor Environmental Metrics Collected During Particle Count^a.

Variable	No. of patients	Mean (SD) [range]
Temperature, °C	76	24.1 (2.1) [18.6-28.2]
Humidity, %
Indoor	76	52.4 (8.3) [35.4-72.2]
Outdoor	79	65.5 (13.3) [35.0-96.0]
No. of particles
>0.5 μm/ft3 of air	79	316 138.5 (753 472.0) [114.4-5 300 000.0]
>5.0 μm/ft3 of air	79	876.9 (1352.8) [0-8843.3]
Mass concentration, μg/m3
PM2.5	83	2.6 (4.8) [0-5.5]
PM10	83	8.0 (9.9) [0-54.0]

Abbreviations: PM_2.5, particulate matter of 2.5 μm or less in aerodynamic diameter; PM₁₀, particulate matter of 10.0 μm or less in aerodynamic diameter.

^aMeasured using a handheld particle counter (Aerocet 531; MetOne Instruments, Inc).

Correlations Between Indoor Environment and Dry Eye Metrics

Exploratory analysis suggests that humidity (r = 0.30 [95% CI, 0.07-0.49]; P = .01) and the number of particles greater than 5.0 μm/ft³ (r = 0.26 [95 CI, 0.03-0.47]; P = .03) were correlated most closely with dry eye symptom severity as measured by OSDI (Table 3). Among the indoor environmental metrics, humidity was most closely correlated with dry eye signs, including a positive correlation with inflammation (r = 0.32 [95% CI = 0.10-0.51]; P = .01), eyelid vascularity (r = 0.27 [95% CI, 0.05-0.47]; P = .02), and meibomian gland dropout (r = 0.27 [95% CI, 0.05-0.47]; P = .02) and negative correlation with Schirmer scores (r = −0.25 [95% CI, −0.45 to 0.02]; P = .03). Overall, these results suggest that humidity is positively associated with both dry eye symptoms and signs. Seasonality was not found to be associated with dry eye symptoms and signs.

Table 3. Correlational Coefficient Between Indoor Environmental and Dry Eye Metrics Taken During Aerocet Count Measurement^a.

Metric	Temperature, oC				Humidity, %				No. of particles >0.5 μm/ft3 of air, loge				No. of particles >5 μm/ft3 of air, loge
	Pearson r	P value	Spearman ρ	P value	Pearson r	P value	Spearman ρ	P value	Pearson r	P value	Spearman ρ	P value	Pearson r	P value	Spearman ρ	P value
Symptoms
DEQ-5	−0.06	.64	−0.12	.33	0.11	.34	0.12	.30	−0.03	.76	−0.01	.92	0.19	.11	0.14	.25
OSDI	0.21	.07	0.15	.22	0.30	.01	0.23	.05	0.13	.27	0.05	.70	0.26	.03	0.25	.04
Mean pain intensity in 1 wk	0.01	.96	−0.07	.56	0.20	.09	0.21	.08	0.03	.79	−0.03	.81	0.17	.15	0.13	.30
NPSI-E	0.05	.67	−0.03	.82	0.17	.14	0.20	.10	−0.03	.83	−0.19	.10	0.14	.24	0.06	.62
Evoked pain to wind	−0.09	.43	−0.10	.41	0.05	.67	0.06	.64	−0.20	.07	−0.24	.04	0.07	.57	0.04	.77
Evoked pain to light	0.04	.71	0.06	.61	0.10	.39	0.15	.22	−0.10	.37	−0.19	.11	0.04	.77	−0.08	.52
Signsb
Osmolarity	0.08	.49	0.02	.86	−0.05	.69	−0.13	.26	0.02	.84	0.05	.65	0.09	.44	0.24	.05
Inflammation	−0.01	.95	−0.04	.74	0.32	.01	0.34	<.001	0.03	.79	0.03	.78	−0.04	.76	−0.15	.21
Tear breakup time	0.03	.82	−0.07	.56	0.16	.16	0.14	.24	−0.18	.11	−0.18	.13	0.02	.89	−0.03	.81
Corneal staining	0.04	.75	0.08	.52	0.11	.34	0.08	.48	0.06	.63	−0.04	.72	0.08	.50	0.08	.52
Schirmer score	−0.20	.08	−0.24	.04	−0.25	.03	−0.26	.03	−0.02	.87	0.01	.90	0.04	.76	0.10	.43
Eyelid vascularity	0.17	.14	0.21	.08	0.27	.02	0.25	.03	0.01	.94	0.01	.91	−0.08	.49	−0.14	.25
Meibomian gland dropout	0.11	.35	0.18	.13	0.27	.02	0.35	<.001	−0.15	.20	−0.14	.24	−0.06	.61	−0.28	.02
Meibum quality	0.11	.35	0.12	.30	0.12	.30	0.14	.23	−0.03	.82	0.03	.80	−0.06	.61	−0.01	.96

Abbreviations: DEQ-5, Dry Eye Questionnaire 5; NPSI-E, Neuropathic Pain Symptom Inventory Modified for the Eye; OSDI, Ocular Surface Disease Index.

^aP values were not adjusted for multiple analyses.

^bValues are from the more severely affected eye with higher values for osmolarity, staining, eyelid parameters, and inflammation and lower tear breakup time and Schirmer score.

Multivariate Models Adjusted for Confounders

Forward linear stepwise regression models that included demographics (eg, age, sex, race, and ethnicity), comorbidities (eg, depression, arthritis, and type 1 or 2 diabetes), and instrumented particulate matter were run separately for PM_2.5 (Table 4) and PM₁₀ (Table 5). In these models, both PM_2.5 and PM₁₀ showed associations with OSDI and 3 signs of dry eye, namely inflammation, Schirmer score, and meibomian gland dropout. For example, a 1-unit increase in the instrumented PM_2.5 was associated with a 1.59 increase in the OSDI score (β = 1.59 [95% CI, 0.58-2.59]; P < .002). Two other covariates, depression (β = 11.38 [95% CI, 1.35-21.41]; P = .03) and arthritis (β = 14.69 [95% CI, 5.19-24.20]; P = .003), showed an association with OSDI. The instrumented PM_2.5 also emerged as associated with signs of dry eye, including Schirmer score (β = −0.39 [95% CI, −0.75 to −0.03]; P = .04), meibomian gland dropout (β = 0.07 [95% CI, 0.01-0.13]; P = .03), and inflammation (β = 0.06 [95% CI, 0.02-0.11]; P = .009). Two other covariates, age (β = 0.03 [95% CI, 0.01-0.05]; P = .002) and body weight (β = 0.01 [95% CI, 0.00-0.01]; P = .04), showed association with inflammation; age showed an association with meibomian gland dropout (β = 0.03 [95% CI, 0.00-0.05]; P = .04); and hypercholesteremia was inversely associated with Schirmer score (β = −3.64 [95% CI, −7.03 to −0.26]; P = .04) (Table 4). A similar trend was observed for the instrumented PM₁₀ (Table 5). However, the association for PM₁₀ was weaker than that for PM_2.5 for OSDI (β = 1.17 [95% CI, 0.43-1.90]; P = .002) and the 3 signs of dry eye (inflammation: β = 0.05 [95% CI, 0.01-0.08; P < .008]; Schirmer score: β = −0.29 [95% CI, −0.55 to −0.02; P = .04]; meibomian gland dropout: β = 0.05 [95% CI, 0.00-0.09]; P = .02]). For example, the effect of a 1-μg/m³ increase in the instrumented PM_2.5 was associated with a 0.39-mm decline in Schirmer score as against a 0.29 decline in the Schirmer score for the instrumented PM₁₀.

Table 4. Forward Stepwise Multivariate Analysis of the Association of Indoor PM_2.5 With Dry Eye Metrics^a .

Variable	OSDI	Dry eye signs
		Inflammation	Schirmer score	Meibomian gland dropout
PM2.5, μg/m3b	1.59 (0.58 to 2.59)c	0.06 (0.02 to 0.11)c	−0.39 (−0.75 to −0.03)d	0.07 (0.01 to 0.13)d
Depression	11.38 (1.35 to 21.41)d	NA	NA	NA
Arthritis	14.69 (5.19 to 24.20)c	NA	NA	NA
Weight	NA	0.01 (0 to 0.01)d	NA	NA
Age	NA	0.03 (0.01 to 0.05)c	NA	0.03 (0 to 0.05)d
Hypercholesteremia	NA	NA	−3.64 (−7.03 to −0.26)d	NA
Constant	9.49 (−0.21 to 19.18)e	−1.74 (−3.47 to −0.01)d	16.72 (13.93 to 19.52)c	−0.24 (−1.79 to 1.32)
No. of observations	70	71	71	71
Total variability explained, R2	0.29	0.21	0.12	0.13

Abbreviations: NA, not applicable; OSDI, Ocular Surface Disease Index; PM_2.5, airborne particulate matter of 2.5 μm or less in aerodynamic diameter.

^aUnless otherwise indicated, data are expressed as β regression coefficient (95% CI).

^bPM_2.5 = PM_2.5 instrumented on RH [relative humidity] = β × RH.

^cP < .01, not adjusted for multiple analyses.

^dP < .05, not adjusted for multiple analyses.

^eP < .10, not adjusted for multiple analyses.

Table 5. Forward Stepwise Multivariate Analysis of the Association of Indoor PM₁₀ With Dry Eye Metrics^a.

Variable	OSDI	Dry eye signs
		Inflammation	Schirmer score	Meibomian gland dropout
PM10, μg/m3b	1.17 (0.43 to 1.90)c	0.05 (0.01 to 0.08)c	−0.29 (−0.55 to −0.02)d	0.05 (0.00 to 0.09)d
Depression	11.38 (1.35 to 21.41)d	NA	NA	NA
Arthritis	14.69 (5.19 to 24.20)c	NA	NA	NA
Weight	NA	0.01 (0 to 0.01)d	NA	NA
Age	NA	0.03 (0.01 to 0.05)c	NA	0.03 (0 to 0.05)d
Hypercholesteremia	NA	NA	−3.64 (−7.03 to −0.26)d	NA
Constant	4.62 (−6.62 to 15.86)	−1.93 (−3.68 to −0.18)d	17.92 (14.43 to 21.41)c	−0.44 (−2.02 to 1.14)
No. of observations	70	71	71	71
Total variability explained, R2	0.29	0.21	0.12	0.13

Abbreviations: NA, not applicable; OSDI, Ocular Surface Disease Index; PM₁₀, airborne particulate matter of 10.0 μm or less in aerodynamic diameter.

^aUnless otherwise indicated, data are expressed as β regression coefficient (95% CI).

^bPM₁₀ = PM₁₀ instrumented on RH [relative humidity] = β × RH.

^cP < .01, not adjusted for multiple analyses.

^dP < .05, not adjusted for multiple analyses.

Discussion

This study found that indoor particulate matter levels were associated with dry eye metrics when instrumented on humidity. Higher particulate matter levels correlated with increased dry eye symptom severity (as measured by OSDI), ocular surface inflammation (as measured by InflammaDry), and meibomian gland dropout. Higher particulate matter levels were also associated with decreased Schirmer scores. These results substantiate the findings of previous epidemiologic studies.^22,23 For example, in a study of 520 healthy individuals in India,²⁴ increased commuting time in highly polluted areas, a surrogate measure of high PM₁₀ levels, was correlated with increased dry eye symptoms (redness, watering, irritation, strain, blurring, and photophobia) and reduced TBUT and Schirmer scores. In another study in which 71 healthy Brazilian taxi drivers carried individual particulate matter samplers both indoors and outdoors,²² total PM_2.5 levels correlated with dry eye signs, including reduced TBUT and tear osmolarity levels, but not symptoms. However, such findings have not been consistent across all studies. In a study of 16 824 individuals in South Korea,²⁵ PM₁₀ levels measured at ambient monitoring stations did not show an association with self-reported dry eye symptoms or a dry eye diagnosis. Furthermore, a study of 396 US office workers found no improvement in dry eye symptoms after high-efficacy particle filters were installed in the office environment and reduced particle levels by 94%.²⁶

In our study, high humidity was also positively associated with various dry eye symptoms and signs. This finding is in contrast to several studies in which low humidity was a risk factor for a dry eye diagnosis.^25,27,28,29 For example, in a study of 16 824 individuals in South Korea,²⁵ a 5% increase in humidity levels was associated with decreased presence of dry eye symptoms and diagnosis (odds ratios, 0.87 and 0.86, respectively). Controlled environmental chambers have also demonstrated a relationship between low humidity and dry eye symptoms and signs in healthy individuals and those with dry eye.^27,28 As relative humidity decreases, the speed of aqueous tear evaporation increases.²⁹ However, higher humidity can also increase the time that particulate matter stays airborne, as well as increase particulate matter mass and size.³⁰ Hygroscopic particulate matter, such as sulfate and nitrate, can absorb water and increase in size under conditions of high humidity.³¹ Also, high humidity can facilitate the production and aerosolization of bioaerosols, such as mold spores and endotoxins, which have been linked to ocular surface inflammation.^8,32,33 We hypothesize that high humidity was positively correlated with various dry eye symptoms and signs in this study owing to the interaction between humidity and particulate matter, rather than as a direct effect of humidity on the eye. As such, it is important to interpret these findings in the context of different climate zones and seasons, both of which influence humidity and airborne particulate matter levels.

It is biologically plausible that particulate matter exposure can be negatively associated with ocular symptoms and signs of dry eye. In a mouse model, topical PM_2.5 exposure increased corneal fluorescein staining and decreased Schirmer scores compared with nonexposed animals.³⁴ Exposed mice also exhibited more corneal epithelial apoptosis, more disorganized conjunctival epithelial layers, and fewer conjunctival goblet cells than nonexposed mice. Furthermore, topical PM_2.5 exposure increased levels of the inflammatory markers tumor necrosis factor and nuclear factor–κB in the cornea. Taken together, these results indicate that topical PM_2.5 exposure impairs tear film stability and induces an inflammatory response such as that seen in some subtypes of dry eye in humans. Topical PM₁₀ exposure in mice demonstrated similar findings.³⁵ Topical particulate matter can also potentiate ocular surface disease. In 1 study, titanium dioxide nanoparticles were used as surrogates of particulate matter exposure in rats exposed to scopolamine with desiccating stress (dry eye model) compared with controls. Decreased corneal clarity and increased inflammatory cell infiltration were seen in both groups. However, the effect of titanium dioxide was greater in rats that received concomitant desiccating stress compared with controls. These results suggest that compromised eyes are more vulnerable to particulate matter exposure than healthy eyes.³⁶

Limitations

As with all studies, our findings should be considered in the context of study limitations that include a defined veteran population. As such, exposures, demographic characteristics, and comorbidities may not be generalizable to other groups. In addition, although the particle counter recorded particle number and humidity, our analysis did not take into account other gaseous air pollutants, such as carbon monoxide and ozone, and particle composition and type, such as heavy metals and bioaerosols (ie, mold and pollen), all of which can affect the ocular surface synergistically as well as exacerbate the effects of particulate matter. Finally, these associations do not necessarily implicate a cause-and-effect relationship between aspects of the indoor environment, such as humidity, and symptoms and signs of dry eye.

Conclusions

This cross-sectional study suggested via direct measurements that the indoor environment was associated with dry eye. Furthermore, this finding emerged in an area with a subtropical climate that has not been extensively studied in relation to indoor and outdoor environmental conditions. Unlike with outdoor air pollution, individuals can improve their indoor environment by regulating humidity, temperature, and airborne particulate matter. This offers a potential therapeutic option to improve and manage dry eye symptoms and signs, although this study does not provide direct cause-and-effect evidence that such options would result in dry eye improvement. In fact, interventions such as improved ventilation and high-efficiency particulate air filters have been shown to decrease asthma symptoms and asthma-related hospital visits.³⁷ These results suggest that future studies might investigate the efficacy of interventions on the indoor environment as they relate to improvements in dry eye signs and symptoms.